Critique and opportunity of artificial intelligence

By 2030, it is estimated that the contribution of artificial intelligence (IA) to the world economy will be 15.7 billion dollars. La IA ha irruptido en sectors muy diversos, desde la industria pharmaceutica hasta la sanidad, la fabricationa, la logística y la agriculture de precision; reducing costs, increasing precision, productivity, growth and innovation, both in the public sector and in the private sector, contributing to environmental sustainability.

Su futuro se orientamente principally hacia tres vertientes: la inteligencia artificial generative; es decir, la IA que, por medio de algoritos que utilizan contenido ya existente en textos, archivos de audio o imágenes, crea nuevos contenidos —esta es, según el Massachusetts Institute of Technology (MIT), una de las tendencias más prometadoras—; artificial intelligence engineeringque tiene un papel importante para que la eficacia de las solutions de IA perdure en el tiempo y pueda actualizarse, y, por ultimo, la inteligencia artificial para la toma de decisionsque tiene la capacidad de modelar las decisions de forma repetible, rastreable y evaluable.

There is no doubt that the benefits of this technology will have important repercussions for those who have the vision to invest in skills, computer capacity and identification of the needs of their sector, but also existen riesgos para quienes no tengan en cuenta el respeto a los derechos humanos, la democracia y los valores éticos en sus políticas a la hora de desarrollar y utilizar la IA. De hecho, la historia reciente abunda en ejemplos de ello.

On March 7, 2022, the article was published in the journal Nature Machine Intelligence Dual use of artificial intelligence powered drug discovery en el que se explicata cómo se invirtió un generador comercial de moleculas impulsado por modelos de aprendizaje automático, que normally penalizaba la toxicity y premiaba la actividad correcta, para premiar, con la nueva lógica, la toxicity. The IA was trained, therefore, using compounds like the nerve agent VX, one of the most toxic chemical warfare agents developed during the 20th century. In less than 6 hours since the start of the internal server, the model generated 40,000 toxic molecules.

In the case of Amazon in 2015, although the IA system was well structured, discriminaba a las mujeres because he had been trained with data from the CVs selected by the company in the last 10 years and which, in most cases, belonged to men.

It is not strange that there are problems in the development of algorithms, as in the case of the COMPAS algorithm used by the judges of a county in Florida to generate a puntuación —baja, media o alta— associated with the risk that a person could recidivism in the siguientes dos años desde su puesta en libertad. Un grupo de periodistas comparó miles de datos y verificó que había un número disproportionado de valoraciones altas en el grupo de personas negras. Pese a ser clasificadas como personas de alto riesgo, no llegaron a commiter ningún delito en los dos años siguientes. Another case is that of the Apple Card, whose algorithm defines credit limits up to 20 times more harmful for female clients with regard to men, even within the same family nucleus.

Todos estos ejemplos ponen de manifesto cómo the absence of an analysis of human rights in the design of Artificial Intelligence puede tener un impacto negativo en su uso, difusión y, en consequencia, en su valor. It is time that Italy orientates the development of technologies, and in particular of artificial intelligence, to put human beings in the center. An example of this is the participation of the Ministry of Foreign Affairs and International Cooperation in the development of the Recommendation on the Ethics of Artificial Intelligence, adopted unanimously by the 41st Session of the General Conference of UNESCO at the end of 2021. Y es gracias a la contribution de dicho Ministerio que ahora se prohíbe de forma explicita la vigilancia massiva en las recommendaciones: “AI systems should not be used for social scoring or mass surveillance purposes”.

The application of this Recommendation queda bajo la responsabilidad de los Estados membersque deben informar al respecto a la UNESCO cada cuatro años.

Sobre esta base, Italia, junto con otros Estados miembros, se unió a la creación de una Red de “Early Adopters/Supporters”, later denominated “Group of Friends”, cuyo objetivo es ayudar a la Secretariat de la UNESCO actuando como foro para compartir las mejores prácticas y debatir los diferentes enfoques y metodologías relacionados con el desarrollo de la IA. This will be combined with an evaluation of the implementation of the Recommendation on the part of UNESCO members (“readiness assessment”), and with skills development activities in countries that do not yet have legislation or knowledge in this area.

Cuando fui Minister of Innovation and Digitalization In the second government of Prime Minister Giuseppe Conte, we are not only working to set the foundations for the use of artificial intelligence in public services — with the obligation to share data between public administrations and create a platform to correlate and analyze data — también para basar el uso de la IA en los valores éticos de transparency y respeto al ser humano que son propios de la Constitución italiana, y en los valores en los que se basa la Unión Europea.

We have begun to establish a group of useful indicators to elect service providers that use AI for public administration with an evaluative balance not only from the economic and technological point of view, but also from the ethical point of view. En nombre del Gobierno italiano, I had the honor of signing an ethical charter, the “Rome call for AI Ethics”, promoted by the Accademia Pontificia per la Vita with the objective of involucar a los estados, las empresas y los individuos interesados ​​en la questión de la ética y la IA, y promover el desarrollo de algoritos transparentes, evaluables e inclusivos en los que la tecnología serva de apoyo para los humanos y no los sustituya. Also, during Conte’s second legislature, we support Italy’s participation in projects such as the Global Partnership for AI (GPAI) where, together with the EU and other 23 countries, Italy works to bridge the gap between the theory provided by the different documents of the OECD (OECD Principles on AI, OECD Recommendation of the Council on Artificial Intelligence y las actividades del OECD Observatory on AI) and practice.

Finally, in November 2021, Italy approved the Strategic Program for Artificial Intelligence 2022-2024 for reinforce las competencias y atraer talento, aumentar la financing de la investigación avanzada en IA and impulsar la adopción de la IA y sus applications, tanto en la Administración Pública (AP) como en los sectors productiveos en general. Italy’s focus on AI is inscribed in a broader framework of the EU’s intention to design “a resilient Europe for the digital decade” in which many people and companies can enjoy the benefits of AI.

Son dos los amítos: la excellence en la IA y la reliability de la IA. The European Commission has developed an AI strategy to accompany Europe’s approach to AI, and published, in April 2021, an AI package proposing new norms and actions to convert Europe into a trusted world center for Artificial Intelligence. El Consejo de Europa no es menos importante, y bajo la presidencia italiana del Comité de Ministers del Consejo de Europe, está centrándose en el desarrollo de sistemas de Inligencia Artificial que se centran en los derechos humanos y democraticos y el Estado de derecho. The Committee of Artificial Intelligence (CAI) received the mandate of the Committee of Ministers to “develop a legal framework appropriate for the development, design and application of artificial intelligence systems”.

The legislative power must also guarantee complementarity with the international norms in force in the matter of human rights, democracy and the State of law, and will have the essential objective colmar las lagunas jurídicas que puedan haber surgido deboto a los rápidos avances tecnológicos en este amíto. This effort will be complementary to the negotiations already initiated in the EU.

The efforts of Italy and the EU on the question of IA and the respect of human rights, ethics and democracy show the intention to eliminate one of the major misunderstandings of the technological development related to rights: put the man en el centro de la tecnología no significa frenar su desarrollo, sino todo lo contrario, significa aumentar su valor. No one will trust totally in a new technology that can have negative consequences for himself or his children in terms of limiting civil, political, economic, social and cultural rights. La falta de confianza solo puede limitar el progresso tecnológico, y la Europe actual no puede permitírselo. No solo por sí misma, sino también por los equilibrios geopoliticos que está encomendada a mantener tanto online como offline. Tenemos que mirar hacia nuestro futuro con valentía y dejarnos guiar, en cada elección, por nuestros valores democraticos.

Paola Pisano is ex-minister of Technological Innovation and Digital Transition of Italy.

.

Leave a Comment

Your email address will not be published. Required fields are marked *